Stimulated by an agonistic ligand, α-galactosylceramide (αGalCer), invariant NKT (iNKT) cells are capable of both eliciting antitumor responses and suppressing autoimmunity, while they become anergic after an initial phase of activation. It is unknown how iNKT cells act as either activators or regulators in different settings of cellular immunity. We examined effects of αGalCer administration on autoimmune inflammation and characterized phenotypes and functional status of iNKT cells and dendritic cells in αGalCer-treated NOD mice. Although iNKT cells became and remained anergic after the initial exposure to their ligand, anergic iNKT cells induce noninflammatory DCs in response to αGalCer restimulation, whereas activated iNKT cells induce immunogenic maturation of DCs in a small time window after the priming. Induction of noninflammatory DCs results in the activation and expansion of islet-specific T cells with diminished proinflammatory cytokine production. The noninflammatory DCs function at inflammation sites in an Ag-specific fashion, and the persistence of noninflammatory DCs critically inhibits autoimmune pathogenesis in NOD mice. Anergic differentiation is a regulatory event that enables iNKT cells to transform from promoters to suppressors, down-regulating the ongoing inflammatory responses, similar to other regulatory T cells, through a ligand-dependent mechanism.
Ligand engagement drives cellular immunity. T cells are activated as such, and then die or become memory cells. For autoreactive T cells, their antigenic responses cause tissue damage, or in the case of FoxP3+ regulatory T cells, the ligand engagement empowers them to suppress the activation of other T cells sharing identical antigenic specificity, as well as bystander responses (1, 2, 3). However, the effects and consequences of activation of invariant NKT (iNKT)3 cells that express Vα14TCR and recognize both endogenous and exogenous glycolipid Ags are ambiguous (4, 5, 6). Upon TCR ligation, iNKT cells release a burst of Th1 and Th2 cytokines, resulting in transactivation of other immune cells, including the maturation of dendritic cells (DCs) that are highly immunogenic and capable of eliciting antitumor responses (7, 8). On the other hand, administration of αGalCer (α-galactosylceramide), a potent agonistic ligand to iNKT cells, has been applied to animal models for different autoimmune diseases, including NOD mice. The treatment prevents, but sometimes fuels, autoimmunity depending on the protocols and genetic backgrounds (9, 10). Therefore, the activated iNKT cells either foster or suppress T cell immunity in different environments, resulting in less predictable results that vastly reduce the potential clinical value of αGalCer in patients with heterogeneous backgrounds.
In addition to the double-role enigma, it is intriguing how iNKT cells act as activators or regulators once they become “anergic” after initial TCR engagement. An αGalCer injection triggers a transient down-regulation of Vα14TCR, followed by a repopulation and expansion of iNKT cells (11). However, cytokine production by the repopulated iNKT cells is severely reduced in response to αGalCer restimulation (12, 13). Since only multiple doses of αGalCer are effective for treatment of autoimmune diseases (14, 15), it is not known whether iNKT cells maintain their unresponsiveness in these mice and whether anergic iNKT cells have any functional effect on subsequent immune responses.
Like other organ-specific autoimmune diseases, a key pathogenic event in type 1 diabetes (T1D) is the local inflammatory responses mediated by autoreactive T cells against pancreatic islets. The activation of these T cells driven by islet Ag-bearing DCs in pancreatic lymph nodes results in the expansion of pathogenic T cell populations that secrete proinflammatory cytokines/chemokines, leading to up-regulation of death receptors/ligands and the elimination of insulin-producing β-cells (16, 17, 18, 19, 20). The antiinflammatory Th2 deviation of autoreactive T cells promoted by iNKT cells was proposed as a protective mechanism in αGalCer-treated NOD mice (14, 21, 22), although regulatory mechanisms independent of Th2 cytokines were also suggested (23, 24). It remains unknown, however, whether and how these mechanisms are responsible for the inhibition of islet-specific inflammation and pathogenesis of T1D.
We show that iNKT cells become and remain hyporesponsive following initial activation. However, in response to αGalCer restimulation, anergic iNKT cells induce noninflammatory DCs that, in turn, inhibit effector differentiation of activated T cells. Furthermore, the persistence of noninflammatory DCs by multiple doses of αGalCer is crucial for the suppression of chronic islet-specific autoimmune inflammation in NOD mice. Therefore, iNKT cells transform from activators to suppressors that share an anergic phenotype with other well-defined regulatory T cells through a ligand-dependent mechanism. This mechanism may balance the role of iNKT cells in effective response to infections and maintenance of self-tolerance, and may have significant clinical implications.
Materials and Methods
NOD, NOD.scid, NOD.Tcrα−/−, NOD.Il10−/−, and NOD.Ifng−/− mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Stat6−/− and Stat6/Il10−/− NOD mice were as described (25). BDC2.5 NOD transgenic mice were from Dr. J. D. Katz (University of Cincinnati, Cincinnati, OH), and 8.3 NOD transgenic mice were from Dr. Pere Santamaria (University of Calgary). All mice were maintained in a specific pathogen-free facility at the University of Calgary, according to the Institutional Animal Care and Use Committee guidelines.
αGalCer was provided by Kirin Brewery (Gunma, Japan). CD1d tetramers were provided by the National Institutes of Health tetramer center. Abs for FACS analysis and intracellular staining were all purchased from BD Pharmingen Canada. DuoSet kits for ELISA were from R&D System.
Mice were injected twice per week i.p. with αGalCer (5 μg/dose) or vehicle for durations as indicated in each experiment. Mice were bled and sacrificed 4 h after the last injection to measure serum cytokines and to harvest cells.
Cell preparation and T cell activation
CD11c+ DCs, T cells, and CD4+ BDC2.5 and CD8+ 8.3 T cells were isolated from wild-type or transgenic donor mice with antibody-conjugated microbeads (Miltenyi Biotec). Peptides P1040-51 and NRP-A7 were used as ligands for in vitro activation of BDC2.5 CD4+ and 8.3 CD8+ T cells, respectively (26, 27). CD4+ T cells (0.5 × 106/ml) were cultured for 3 days with various concentrations of immobilized anti-CD3 Ab, and CD4+ BDC2.5 or CD8+ 8.3 T cells (0.5 × 106/ml) were cultured with DCs (0.1 × 106/ml) for 3 days in the presence of various concentrations of ligands. [3H]thymidine incorporation was used to measure T cell proliferation, and cytokine production was measured using ELISA.
Th1.2+ T cells and total or DC-depleted splenocytes were transferred into 7–8-wk-old NOD.tcrα−/− or NOD.scid recipients (10 × 106 cells/recipient), and the development of T1D in the recipients was monitored twice weekly. For cotransfer experiments, DC-depleted splenocytes from diabetic NOD mice (10 × 106 cells/recipient) were mixed with DCs (2 × 106 cells/recipient) from different donor mice with or without islet Ag loading. For the islet Ag loading, DCs were incubated with freeze-thaw-disrupted islets for 3 h before being used for cell transfer. Additional DC transfers were performed every 3 wk as indicated. Purified BDC2.5 T cells were labeled with 10 mM CFSE and transferred i.v. into control and αGalCer-treated mice (8 × 106 cells/recipient).
Inguinal lymph nodes (ILNs), mesenteric lymph nodes (MLNs), and pancreatic lymph nodes (PLNs) were removed from control and αGalCer-treated NOD mice. Lymph node cells were resuspended in RPMI 1640 culture medium supplemented with 10% FCS and containing 2.0 μM monesin (GolgiStop, BD Phamingen), and stimulated with PMA (10 ng/ml) and calcium ionophore (250 ng/ml) at 37°C for 6 h. Cells were then permeabilized and fixed and stained with anti-IFN-γ-PE Ab. Cells were further stained with anti-TCR-PerCP Ab and analyzed by FACS. Data were analyzed using FlowJo program (TreeStar).
All statistical analyses were performed with GraphPad Prism 5 software. The results of in vivo studies of adoptive cell transfer were analyzed with a Kaplan-Meier log-rank test. For in vitro and intracellular staining studies, differences between two groups were evaluated with an unpaired Student’s t test or Mann-Whitney U test, and the differences among groups were analyzed with one-way or two-way ANOVA followed by a Newman-Kuel post hoc test or Dunnett’s test as appropriate. In all cases, we considered p < 0.05 as statistically significant.
DCs from αGalCer-treated NOD mice reduced IFN-γ production by islet-specific T cells
To assess the effects of αGalCer administration on islet-specific autoimmune inflammation, we examined IFN-γ response in PLNs of NOD mice. This IFN-γ response was islet specific, since IFN-γ+ T cells were hardly detectable in ILNs of prediabetic female NOD mice. The frequencies of IFN-γ+ T cells reflected the intensity of inflammation, as they were significantly increased in PLNs and even ILNs of newly diabetic mice (Fig. 1,a). However, islet-specific IFN-γ response was severely reduced in NOD mice treated with multiple doses of αGalCer (Fig. 1 b). In contrast, a singe dose of αGalCer did not affect this response.
To assess the effects of αGalCer treatment on the functions of T cells and DCs, we isolated CD4+ T cells and CD11c+ DCs from control and multiple dose αGalCer-treated NOD mice. CD4+ T cells from the αGalCer-treated mice produced more IL-4 but similar levels of IFN-γ than did those from control mice when activated by anti-CD3/CD28 Abs (Fig. 1,c), suggesting a Th2-orientated immune deviation. On the other hand, both CD4+ BDC2.5 and CD8+ 8.3 T cells, two representative clonal islet-specific T cells from TCR transgenic NOD mice (28, 29), produced markedly reduced IFN-γ when responding to their ligands (peptide 1040–51 and NRP-A7 (26, 27), respectively) presented by DCs from the mice treated with multiple doses of αGalCer (M-DCs) than by those from control (C-DCs) mice, although these T cells proliferated equally well (Fig. 1, d and e). αGalCer can potently induce maturation of DCs (30); however, M-DCs seemed less immunogenic than did immature C-DCs.
Noninflammatory DCs induced by αGalCer inhibited diabetogenic T cells
To investigate the effects of immune deviation on T1D pathogenesis, T cells isolated from control and αGalCer-treated mice were introduced into NOD.Tcrα−/− recipients. T1D developed in every recipient, indicating that T cells in αGalCer-treated NOD mice were still able to differentiate into diabetogenic effectors once released from the suppressive environment, regardless of their Th2-like profile (Fig. 2,a). In contrast to T cells, total splenocytes from αGalCer-treated mice, as reported previously (14, 22), demonstrated a severely reduced ability to induce T1D once transferred into NOD.scid recipients. However, depletion of DCs restored the diabetogenicity of these splenocytes (Fig. 2 b), suggesting a pivotal role of DCs for T1D inhibition.
We then cotransferred C-DCs or M-DCs with DC-depleted splenocytes from newly diabetic NOD mice that alone induced T1D in NOD.scid recipients rapidly. Replacement of DCs in diabetic splenocytes with M-DCs barely slowed the pathogenic progress in the recipients. However, the inhibitory effects of M-DCs were clear when M-DCs were loaded with islet Ags before the transfer. The T1D incidence was significantly reduced in the recipients of diabetic splenocytes with two additional transfers of islet Ag-loaded M-DCs (once every 3 wk) (Fig. 2 c) that might have compensated for M-DCs turnover. In contrast, islet Ag-loaded C-DCs hardly slowed the onset of T1D in the recipients of diabetic splenocytes. These results show that M-DCs, but not C-DCs, inhibited diabetogenic T cells in an Ag-specific fashion.
iNKT cells and DCs were differentiated by different αGalCer treatments
Since αGalCer injection activates but also induces hyporesponsiveness of iNKT cells (12), we characterized the functional status of iNKT cells in response to single and multiple αGalCer stimulations. Because the iNKT cell deficiency associated with autoimmunity in NOD mice (9) might be corrected by αGalCer treatment, we compared iNKT cell activation in both NOD and B6.g7 mice that share an identical MHC locus with NOD mice but had a normal iNKT cell population. A single dose or five doses of αGalCer were injected into NOD and B6.g7 mice, and serum cytokines were measured 4 h after the last αGalCer injection. A cytokine surge was detected after a single dose of αGalCer, although the levels of IFN-γ and IL-4 were much lower in NOD than in B6.g7 mice. However, cytokines were not detected in the serum of either NOD or B6.g7 mice injected with multiple doses of αGalCer (Fig. 3,a). The diminished cytokine production was not due to the iNKT cell deficiency in NOD mice or reduction of the iNKT cell population, since splenic iNKT cells expanded slightly by multiple doses of αGalCer (Fig. 3,b). Furthermore, an additional dose of αGalCer after a 2-wk rest only induced low levels of IL-4 but not IFN-γ, showing an extended period of hyporesponsiveness of iNKT cells by multiple doses of αGalCer (Fig. 3,a). Further ex vivo analyses showed that iNKT cells were hyperresponsive for a short period of time after the αGalCer priming before turning to be hyporesponsive to subsequent stimulations (Fig. 3, c and d).
We then defined the effects of αGalCer treatment on the maturation of splenic DCs. In 4 h, CD40, CD80, and CD86 expressions were up-regulated in CD8+ DCs isolated from mice exposed to a single dose of αGalCer (S-DCs). In contrast, the levels of CD80 and CD86 were down-regulated in M-DCs (Fig. 4,a). The differences were less pronounced in a CD11c+CD8− subset (data not shown). Additionally, S-DCs produced high levels of TNF-α and IL-12, while M-DCs produced less IL-12 than did both S-DCs and C-DCs in response to a TLR9 agonist CpG (Fig. 4 b).
Importantly, DCs matured by single and multiple doses of αGalCer acquired different abilities to activate 8.3 T cells. In comparison to C-DCs, DCs from mice 4 (S-DCs) and 24 h (S1-DCs) after a single αGalCer injection significantly increased IFN-γ production by 8.3 T cells. In contrast, M-DCs reduced IFN-γ production but not proliferation of 8.3 T cells (Fig. 4,c). To confirm the iNKT cell-dependent modulation of DCs, we injected αGalCer into CD1d-deficient NOD mice and found that DCs from αGalCer-injected CD1d-deficient NOD mice did not elicit or regulate IFN-γ production of islet-specific T cells (Fig. 4,d). When BDC2.5 T cells were activated by their ligands presented by C-DCs, S-DCs, and M-DCs, both their IFN-γ and IL-10 productions were driven up by S-DCs but down by M-DCs, arguing against a selective Th polarization (Fig. 4,e). Additionally, IFN-γ and IL-10 productions by BDC2.5 T cells from Stat4-deficient mice were also elicited or reduced by S-DCs or M-DCs (Fig. 4 d and data not shown), indicating an IL-12-independent regulation. Thus, DCs matured by single αGalCer injection were immunogenic, but DCs became noninflammatory after multiple doses of αGalCer. These results imply that iNKT cells regulate autoreactive T cells via DCs.
Induction and maintenance of noninflammatory DCs by anergic iNKT were ligand dependent
To ascertain that iNKT cells from αGalCer-treated NOD mice, activated or anergic, directly instruct the differentiated DC maturation, we cocultured immature splenic C-DCs in the presence or absence of αGalCer with enriched iNKT cells from control (C-iNKT, naive) or multiple dose αGalCer-treated mice (M-iNKT, anergic). After 2 days, DCs were isolated from culture mixes and used as APCs for islet-specific T cells.
Vigorous proliferative responses of islet-specific T cells demonstrated equal abilities as APCs of DCs cocultured with different iNKT cells. However, DCs cocultured with C-iNKT cells in the presence of αGalCer significantly increased IFN-γ production by islet-specific T cells, showing that the activation of naive C-iNKT cells induced immunogenic maturation of C-DCs. In contrast, C-DCs that had been cocultured with M-iNKT cells in the presence of αGalCer reduced IFN-γ production by the islet-specific T cells (Fig. 5, a and b). Therefore, αGalCer-stimulated anergic M-iNKT cells promoted noninflammation differentiation of C-DCs.
Although iNKT cells maintain a long-term anergy after a single dose of αGalCer (12), we found that DCs isolated from NOD mice 2 wk after a single dose of αGalCer (SR-DCs), however, were not tolerogenic, as they expressed CD80 at levels similar or even slightly higher than that of C-DCs, and they also triggered IFN-γ responses by BDC2.5 T cells comparable to C-DCs (Fig. 5, c and d). In contrast, DCs from mice 2 wk after multiple doses of αGalCer (MR-DCs) remained noninflammatory, since they expressed lower levels of CD80 than did C-DCs and induced low levels of IFN-γ by BDC2.5 T cells, indicating the persistence of noninflammatory DCs by multiple doses of αGalCer.
Th2 cytokines and IFN-γ were dispensable for induction of noninflammatory DCs
Because anergic iNKT cells retained their IL-4 capability, we investigated the role of IL-4 as well as other Th2 cytokines in the induction of noninflammatory DCs by injecting αGalCer into female Stat6−/− and Stat6/Il10−/− NOD mice. The development of iNKT cells in NOD.Stat6/Il10−/− mice was similar to NOD mice (data not shown), although the deficient IL-4, IL-10, and IL-13 signaling resulted in accelerated pathogenesis of T1D (25). The treatment of αGalCer did not alter IFN-γ and proliferative responses of CD4+ T cells in response to anti-CD3 stimulation in culture (Fig. 6,a); however, DCs were tolerized in these mice since they significantly reduced IFN-γ production by islet-specific T cells (Fig. 6, b and c). In accordance with the induction of noninflammatory DCs, both Stat6−/− and Stat6/Il10−/− NOD mice were similarly protected as were NOD mice by αGalCer treatment (Fig. 6,d–f). We also found that αGalCer induced immunogenic S-DCs and tolerogenic M-DCs in NOD.Ifng−/− mice and protected these mice from T1D (Fig. 6 g–i). Therefore, neither Th2 cytokines nor IFN-γ by iNKT cells was required for differentiated maturation of DCs.
DCs in PLNs of αGalCer-treated NOD mice were tolerogenic
To inhibit islet-specific autoimmune inflammation, noninflammatory DCs must act in PLNs. We isolated DCs from PLNs, MLNs, and spleen of control and multiple dose αGalCer-treated NOD mice and tested their function as APCs. In control NOD mice, the overall activity of DCs from PLNs was lower than that from the spleen and MLNs, as reported previously (23). Nevertheless, DCs from PLNs, but not from MLNs, of αGalCer-treated mice reduced antigenic IFN-γ response by 8.3 T cells in culture (Fig. 7 a), suggesting that αGalCer administration did not induce systemic tolerance and that noninflammatory DCs selectively functioned at inflammation sites.
We tested the function of DCs in PLNs in vivo by transfusing CFSE-labeled BDC2.5 T cells into control and αGalCer-injected (5 doses) NOD mice. PLNs and ILNs were removed 4 days after the cell transfer, and proliferative and IFN-γ responses of CFSE+ cells were analyzed. As expected, BDC2.5 T cells proliferated in PLNs but not in ILNs, and most CFSE+ cells in PLNs of control mice were IFN-γ+. In contrast, CFSE+ T cells in PLNs from αGalCer-treated NOD mice were not IFN-γ+, although they proliferated vigorously (Fig. 7 b). Therefore, both in vivo and ex vivo, noninflammatory DCs in αGalCer-treated NOD mice demonstrated the ability to activate autoreative T cells with diminished IFN-γ production.
DCs have been increasingly recognized as a key player in the establishment and maintenance of self-tolerance (31, 32, 33, 34). However, DCs in NOD mice, as an essential component of T1D pathogenesis, present both MHC class I- and II-restricted islet Ags to autoreactive T cells (19, 35) to initiate autoreactive responses. The abnormally high proportions of myeloid DCs with dysregualted NF-κB activity in NOD mice (36, 37, 38) may contribute to the failed self-tolerance. Induction of tolerogenic properties of DCs, such as expression of IDO that blocked the expansion of islet-specific T cells, inhibited autoimmune pathogenesis (25, 35). However, in the present study noninflammatory DCs induced by anergic iNKT cells resulted in the proliferation of islet-specific T cells with a reduced pathogenic potential, as indicated by the diminished production of IFN-γ and perhaps other inflammatory cytokines as well. Noninflammatory DCs capable of inhibiting the effector functions of activated islet-specific T cells were crucial in the suppression of T1D pathogenesis, whereas the diabetogenecity retained in T cells from αGalCer-treated NOD mice indicated that anergic iNKT cells did not suppress pathogenic T cells directly.
An early study detected the enrichment of myeloid DCs with reduced IL-12 capacity in PLNs of αGalCer-treated NOD mice, and vaccination of young NOD mice with islet Ag-bearing myeloid DCs from untreated NOD mice through footpad injection reduced T1D by inducing a regulatory component (24, 39). On the other hand, multiple doses of αGalCer promoted IL-10 capacity of splenic DCs in B6 mice (40). Surprisingly, neither the induction nor the function of noninflammatory DCs in NOD mice depended on IL-12, IL-10, or other Th2 cytokines. Exactly how noninflammatory DCs reduce IFN-γ response of islet-specific T cells in PLNs despite the intensified recruitment and expansion of T cells by αGalCer (Ref. 23 and our unpublished data) remains to be characterized. Importantly, noninflammatory DCs minimized autoreactive responses specifically in the inflammatory sites. Splenic DCs in αGalCer-treated NOD mice were tolerogenic, likely because the spleen is a major residential site for iNKT cells where they abundantly interact with DCs (41). It is possible that noninflammatory DCs preferentially migrate, driven by chronic inflammatory signals, from spleen into PLNs, since inflammation enhances lymph node-bound migration of DCs (42). By this mechanism, DCs in other locations, such as MLNs, could maintain their immature status and alertness on infectious challenges.
The induction of noninflammatory DCs was ligand dependent, as anergic iNKT cells did not effectively induce noninflammatory DCs in the absence of αGalCer restimulation either in vitro or in vivo. Although a single dose of αGalCer induced long-term anergy of iNKT cells (12), DCs were not tolerogenic, and the islet-specific inflammation was not affected. Repetitive injections of αGalCer, therefore, were required for the persistence of noninflammatory DCs, which might be crucial for protection of NOD mice in which islet-specific autoreactive inflammation preceded the onset of hyperglycemia for a protracted period, whereas lifespan of DCs was shortened by maturation (43). Supporting this model, extending the presence of noninflammatory DCs in NOD.scid recipients by multiple adoptive transfers inhibited T1D by diabetogenic T cells, whereas the single transfer of noninflammatory DCs had a limited effect. The presence of αGalCer might enhance direct contacts of DCs with anergic iNKT cells, as even a low concentration of αGalCer significantly stabilized the CD1d-dependent interactions (41). Alternatively, products other than Th2 cytokines by αGalCer-stimulated anergic iNKT cells might also modulate the function of DCs.
Consistent with previous reports (8, 44), activated iNKT cells induced immunogenic DCs that elicited IFN-γ by islet-specific T cells. In contrast, anergic iNKT cells induced noninflammatory DCs. These results revealed a physiological significance of the dichotomous activation status of iNKT cells, as well as a mechanism by which iNKT cells promote or regulate cellular immunity. Rapid and robust activation of iNKT cells benefits host defense by initiating immune responses against pathogens or cancer cells through induction of immunogenic DCs. However, the potent bystander responses induced by activated iNKT cells may result in tissue damage (45, 46, 47). Anergic iNKT cells prevent further response cascade when reencountering the ligand through induction of noninflammatory DCs that down-regulate ongoing inflammation once recruited into inflammatory locations. In this process, the anergic differentiation of the activated iNKT cells is a turning point from being a promoter to a regulator. It is likely for this reason that a single dose of αGalCer at the time of immunization potentiated EAE in B10.PL mice, but preinjection of multiple doses of αGalCer inhibited the disease (15). Similarly, islet grafts that were otherwise acutely rejected by the activated iNKT cells were protected by multiple doses of αGalCer (48). These observations support the ligand-dependent regulatory mechanism by anergic iNKT cells. On the other hand, failure in anergic differentiation of iNKT cells may have detrimental effects on self-tolerance as demonstrated in NZB/NZW and SJL mice in which spontaneous or induced lupus was exacerbated by iNKT cells with the elevated IFN-γ in response to multiple doses of αGalCer (10, 49, 50). An implication of this study, therefore, concerns screening the response of iNKT cells from patients to multiple doses of αGalCer to reduce the risk of tissue damages from bystander responses.
The authors have no financial conflicts of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the Juvenile Diabetes Research Foundation International, Canadian Diabetes Association, and the Julia McFarlane Diabetes Research Centre (to Y.Y.).
Abbreviations used in this paper: iNKT, invariant NKT; αGalCer, α-galactosylceramide; DC, dendritic cell; ILN, inguinal lymph nodes; MLN, mesenteric lymph node; PLN, pancreatic lymph node; T1D, type I diabetes.